feat(wiki): support local Claude and Codex providers

gitnexus/scripts/cross-platform-tests.ts

@@ -79,6 +79,7 @@ const SPAWN_CLI = [
   'test/integration/group/group-cli.test.ts',
   'test/integration/cli/tool-no-index-stderr.test.ts',
   'test/integration/setup-skills.test.ts',
+  'test/unit/local-cli-subprocess.test.ts',
 ];
 
 // Worker threads tests — exercise real worker_threads which have

gitnexus/src/cli/i18n/en.ts

@@ -185,7 +185,8 @@ export const en = {
   'help.option.clean.all': 'Clean all indexed repos',
   'help.option.clean.lbugSidecars': 'Clean quarantined LadybugDB missing-shadow WAL sidecars',
   'help.option.wiki.force': 'Force full regeneration even if up to date',
-  'help.option.wiki.provider': 'LLM provider: openai or cursor (default: openai)',
+  'help.option.wiki.provider':
+    'LLM provider: openai, openrouter, azure, custom, cursor, claude, or codex (default: openai)',
   'help.option.wiki.model': 'LLM model or Azure deployment name (default: minimax/minimax-m2.5)',
   'help.option.wiki.baseUrl':
     'LLM API base URL. Azure v1: https://{resource}.openai.azure.com/openai/v1',

gitnexus/src/cli/i18n/zh-CN.ts

@@ -174,7 +174,8 @@ export const zhCN = {
   'help.option.clean.all': '清理所有已索引仓库',
   'help.option.clean.lbugSidecars': '清理已隔离的 LadybugDB missing-shadow WAL sidecar',
   'help.option.wiki.force': '即使已是最新也强制完整重新生成',
-  'help.option.wiki.provider': 'LLM 提供商：openai 或 cursor（默认：openai）',
+  'help.option.wiki.provider':
+    'LLM 提供商：openai、openrouter、azure、custom、cursor、claude 或 codex（默认：openai）',
   'help.option.wiki.model': 'LLM 模型或 Azure deployment 名称（默认：minimax/minimax-m2.5）',
   'help.option.wiki.baseUrl':
     'LLM API base URL。Azure v1：https://{resource}.openai.azure.com/openai/v1',

gitnexus/src/cli/index.ts

@@ -146,7 +146,10 @@ program
   .command('wiki [path]')
   .description('Generate repository wiki from knowledge graph')
   .option('-f, --force', 'Force full regeneration even if up to date')
-  .option('--provider <provider>', 'LLM provider: openai or cursor (default: openai)')
+  .option(
+    '--provider <provider>',
+    'LLM provider: openai, openrouter, azure, custom, cursor, claude, or codex (default: openai)',
+  )
   .option('--model <model>', 'LLM model or Azure deployment name (default: minimax/minimax-m2.5)')
   .option(
     '--base-url <url>',

gitnexus/src/cli/wiki.ts

@@ -19,6 +19,7 @@ import {
 import { WikiGenerator, type WikiOptions } from '../core/wiki/generator.js';
 import { resolveLLMConfig, type LLMProvider } from '../core/wiki/llm-client.js';
 import { detectCursorCLI } from '../core/wiki/cursor-client.js';
+import { detectLocalCLI } from '../core/wiki/local-cli-client.js';
 import { logger } from '../core/logger.js';
 
 export interface WikiCommandOptions {
@@ -55,6 +56,18 @@ function parsePositiveIntegerOption(
   return parsed;
 }
 
+function isLocalProvider(
+  provider: LLMProvider | undefined,
+): provider is 'cursor' | 'claude' | 'codex' {
+  return provider === 'cursor' || provider === 'claude' || provider === 'codex';
+}
+
+function localModelConfigKey(provider: 'cursor' | 'claude' | 'codex') {
+  if (provider === 'cursor') return 'cursorModel';
+  if (provider === 'claude') return 'claudeModel';
+  return 'codexModel';
+}
+
 /**
  * Prompt the user for input via stdin.
  */
@@ -191,10 +204,11 @@ const wikiCommandImpl = async (inputPath?: string, options?: WikiCommandOptions)
     if (options.provider) updates.provider = options.provider;
     if (options.apiVersion) updates.apiVersion = options.apiVersion;
     if (options.reasoningModel !== undefined) updates.isReasoningModel = options.reasoningModel;
-    // Save model to appropriate field based on provider
+    // Save model to appropriate field based on provider.
     if (options.model) {
-      if (options.provider === 'cursor') {
-        updates.cursorModel = options.model;
+      const targetProvider = options.provider ?? existing.provider;
+      if (isLocalProvider(targetProvider)) {
+        updates[localModelConfigKey(targetProvider)] = options.model;
       } else {
         updates.model = options.model;
       }
@@ -205,7 +219,7 @@ const wikiCommandImpl = async (inputPath?: string, options?: WikiCommandOptions)
 
   const savedConfig = await loadCLIConfig();
   const hasSavedConfig = !!(
-    savedConfig.provider === 'cursor' ||
+    isLocalProvider(savedConfig.provider) ||
     (savedConfig.apiKey && savedConfig.baseUrl)
   );
   const hasCLIOverrides = !!(
@@ -231,56 +245,84 @@ const wikiCommandImpl = async (inputPath?: string, options?: WikiCommandOptions)
   if (!hasSavedConfig && !hasCLIOverrides) {
     if (!process.stdin.isTTY) {
       // Non-interactive mode — need either API key or Cursor CLI
-      if (!llmConfig.apiKey && llmConfig.provider !== 'cursor') {
+      if (!llmConfig.apiKey && !isLocalProvider(llmConfig.provider)) {
         console.log('  Error: No LLM API key found.');
         console.log('  Set OPENAI_API_KEY or GITNEXUS_API_KEY environment variable,');
-        console.log('  or pass --api-key <key>, or use --provider cursor.\n');
+        console.log('  or pass --api-key <key>, or use --provider cursor|claude|codex.\n');
         process.exitCode = 1;
         return;
       }
       // Non-interactive with env var or cursor — just use it
     } else {
       console.log("  No LLM configured. Let's set it up.\n");
       console.log(
-        '  Supports OpenAI, OpenRouter, Azure, any OpenAI-compatible API, or Cursor CLI.\n',
+        '  Supports OpenAI, OpenRouter, Azure, any OpenAI-compatible API, Cursor CLI, Claude CLI, or Codex CLI.\n',
       );
 
-      // Check if Cursor CLI is available
+      // Check if local agent CLIs are available.
       const hasCursor = detectCursorCLI();
+      const hasClaude = detectLocalCLI('claude');
+      const hasCodex = detectLocalCLI('codex');
+      const localChoices: Array<{
+        choice: string;
+        provider: 'cursor' | 'claude' | 'codex';
+      }> = [];
 
       // Provider selection
       console.log('  [1] OpenAI (api.openai.com)');
       console.log('  [2] OpenRouter (openrouter.ai)');
       console.log('  [3] Azure OpenAI');
       console.log('  [4] Custom endpoint');
+      let nextChoice = 5;
       if (hasCursor) {
-        console.log('  [5] Cursor CLI (local, uses your Cursor subscription)');
+        const choice = String(nextChoice++);
+        localChoices.push({
+          choice,
+          provider: 'cursor',
+        });
+        console.log(`  [${choice}] Cursor CLI (local, uses your Cursor subscription)`);
+      }
+      if (hasClaude) {
+        const choice = String(nextChoice++);
+        localChoices.push({
+          choice,
+          provider: 'claude',
+        });
+        console.log(`  [${choice}] Claude CLI (local, uses your Claude Code login)`);
+      }
+      if (hasCodex) {
+        const choice = String(nextChoice++);
+        localChoices.push({
+          choice,
+          provider: 'codex',
+        });
+        console.log(`  [${choice}] Codex CLI (local, uses your Codex login)`);
       }
       console.log('');
 
-      const maxChoice = hasCursor ? '5' : '4';
+      const maxChoice = String(nextChoice - 1);
       const choice = await prompt(`  Select provider (1/${maxChoice}): `);
 
       let baseUrl: string;
       let defaultModel: string;
       let provider: LLMProvider = 'openai';
       let key = '';
 
-      if (choice === '5' && hasCursor) {
-        // Cursor CLI selected - model defaults to 'auto' (Cursor's default)
-        provider = 'cursor';
+      const selectedLocal = localChoices.find((item) => item.choice === choice);
+      if (selectedLocal) {
+        // Local CLI selected - model defaults to the CLI's configured default.
+        provider = selectedLocal.provider;
         baseUrl = '';
 
-        const modelInput = await prompt('  Model (leave empty for auto): ');
+        const modelInput = await prompt('  Model (leave empty for CLI default): ');
         const model = modelInput || '';
 
-        // Save config for Cursor
-        const cursorConfig: Record<string, string> = { provider: 'cursor' };
-        if (model) cursorConfig.cursorModel = model;
-        await saveCLIConfig(cursorConfig);
+        const localConfig = { ...savedConfig, provider };
+        if (model) (localConfig as Record<string, unknown>)[localModelConfigKey(provider)] = model;
+        await saveCLIConfig(localConfig);
         console.log('  Config saved to ~/.gitnexus/config.json\n');
 
-        llmConfig = { ...llmConfig, provider: 'cursor', model, apiKey: '', baseUrl: '' };
+        llmConfig = { ...llmConfig, provider, model, apiKey: '', baseUrl: '' };
       } else if (choice === '3') {
         // Azure OpenAI guided setup — minimal prompts
         console.log('\n  Azure OpenAI setup.\n');
@@ -328,6 +370,7 @@ const wikiCommandImpl = async (inputPath?: string, options?: WikiCommandOptions)
         const azureBaseUrl = `${endpoint}/openai/v1`;
 
         await saveCLIConfig({
+          ...savedConfig,
           apiKey: azureKey,
           baseUrl: azureBaseUrl,
           model: deploymentName,

gitnexus/src/core/wiki/generator.ts

@@ -39,6 +39,7 @@ import {
 } from './llm-client.js';
 
 import { callCursorLLM, resolveCursorConfig } from './cursor-client.js';
+import { callClaudeLLM, callCodexLLM, resolveLocalCLIConfig } from './local-cli-client.js';
 
 import {
   GROUPING_SYSTEM_PROMPT,
@@ -203,7 +204,7 @@ export class WikiGenerator {
   }
 
   /**
-   * Route LLM call to the appropriate provider (OpenAI-compatible or Cursor CLI).
+   * Route LLM call to the appropriate provider.
    */
   private async invokeLLM(
     prompt: string,
@@ -217,6 +218,16 @@ export class WikiGenerator {
       });
       return callCursorLLM(prompt, cursorConfig, systemPrompt, options);
     }
+    if (this.llmConfig.provider === 'claude' || this.llmConfig.provider === 'codex') {
+      const localConfig = resolveLocalCLIConfig({
+        model: this.llmConfig.model,
+        workingDirectory: this.repoPath,
+        requestTimeoutMs: this.llmConfig.requestTimeoutMs,
+      });
+      return this.llmConfig.provider === 'claude'
+        ? callClaudeLLM(prompt, localConfig, systemPrompt, options)
+        : callCodexLLM(prompt, localConfig, systemPrompt, options);
+    }
     return callLLM(prompt, this.llmConfig, systemPrompt, options);
   }
 

gitnexus/src/core/wiki/llm-client.ts

@@ -9,7 +9,14 @@ import { CircuitOpenError, ResilientFetchExhaustedError, resilientFetch } from '
  * Config priority: CLI flags > env vars > defaults
  */
 
-export type LLMProvider = 'openai' | 'openrouter' | 'azure' | 'custom' | 'cursor';
+export type LLMProvider =
+  | 'openai'
+  | 'openrouter'
+  | 'azure'
+  | 'custom'
+  | 'cursor'
+  | 'claude'
+  | 'codex';
 
 export interface LLMConfig {
   apiKey: string;
@@ -18,7 +25,7 @@ export interface LLMConfig {
   maxTokens: number;
   temperature: number;
   /** Provider type — controls auth header behaviour */
-  provider?: 'openai' | 'openrouter' | 'azure' | 'custom' | 'cursor';
+  provider?: LLMProvider;
   /** Azure api-version query param (e.g. '2024-10-21'). Appended to URL when set. */
   apiVersion?: string;
   /** When true, strips sampling params and uses max_completion_tokens instead of max_tokens */
@@ -44,6 +51,17 @@ export interface LLMResponse {
 export async function resolveLLMConfig(overrides?: Partial<LLMConfig>): Promise<LLMConfig> {
   const { loadCLIConfig } = await import('../../storage/repo-manager.js');
   const savedConfig = await loadCLIConfig();
+  const savedProvider = overrides?.provider ?? savedConfig.provider;
+  const savedLocalModel =
+    savedProvider === 'cursor'
+      ? savedConfig.cursorModel
+      : savedProvider === 'claude'
+        ? savedConfig.claudeModel
+        : savedProvider === 'codex'
+          ? savedConfig.codexModel
+          : undefined;
+  const localProvider =
+    savedProvider === 'cursor' || savedProvider === 'claude' || savedProvider === 'codex';
 
   const apiKey =
     overrides?.apiKey ||
@@ -61,13 +79,12 @@ export async function resolveLLMConfig(overrides?: Partial<LLMConfig>): Promise<
       'https://openrouter.ai/api/v1',
     model:
       overrides?.model ||
-      process.env.GITNEXUS_MODEL ||
-      (savedConfig.provider === 'cursor' ? savedConfig.cursorModel : undefined) ||
-      savedConfig.model ||
-      'minimax/minimax-m2.5',
+      (localProvider ? undefined : process.env.GITNEXUS_MODEL) ||
+      savedLocalModel ||
+      (localProvider ? '' : savedConfig.model || 'minimax/minimax-m2.5'),
     maxTokens: overrides?.maxTokens ?? 16_384,
     temperature: overrides?.temperature ?? 0,
-    provider: overrides?.provider ?? savedConfig.provider ?? 'openai',
+    provider: savedProvider ?? 'openai',
     apiVersion:
       overrides?.apiVersion || process.env.GITNEXUS_AZURE_API_VERSION || savedConfig.apiVersion,
     isReasoningModel: overrides?.isReasoningModel ?? savedConfig.isReasoningModel,